Occurrence and Correlations Between Coliphages and Anthropogenic Viruses in the Massachusetts Bay Using Enrichment and ICC-Npcr Nicola A

59 Q IWA Publishing 2005 Journal of Water and Health | 03.1 | 2005

Occurrence and correlations between coliphages and anthropogenic viruses in the Massachusetts Bay using enrichment and ICC-nPCR Nicola A. Ballester, Justin H. Fontaine and Aaron B. Margolin

ABSTRACT

We evaluated a two-step enrichment procedure to detect coliphages and an integrated cell culture- Nicola A. Ballester (corresponding author) Justin H. Fontaine nested polymerase chain reaction (ICC-nPCR) to detect human astrovirus, enteroviruses, rotavirus Aaron B. Margolin and adenovirus type 40 and 41 in marine water samples collected by the Massachusetts Water Department of Microbiology, University of New Hampshire, Resource Authority (MWRA). MWRA has been monitoring its receiving waters for coliphages, 35 Colovos Rd, ETB Hall Rm. 230, Durham, anthropogenic viruses and indicator bacteria in order to evaluate the impact of Boston’s Deer Island NH 03824, USA Sewage Treatment Plant discharge. Coliphages and enteric viruses were originally assayed using Tel: (603) 862-1172 single agar overlay and most probable number cell culture (MPN) methods, respectively. Reanalysis of Fax: (603) 862-3957 E-mail: [email protected] these samples for enteric viruses by ICC-nPCR demonstrated that 46% were positive for at least one virus compared with 23% with the MPN method. Use of the enrichment method showed a 47% increase in the detection of male specific and somatic coliphages compared with the single agar overlay method. Correlations between the presence of coliphages, enteric viruses and indicator bacteria were based on proximity to the treatment plant discharge, seasonal variations and site levels. The presence of enteric viruses was significantly correlated to coliphages but not to indicator bacteria. Preliminary comparative results demonstrate that effective and efficient monitoring of anthropogenic contamination can be achieved using these more sensitive and specific techniques. Key words | adenovirus, astrovirus, coliphages, enterovirus, ICC-nPCR, rotavirus

INTRODUCTION

Beginning in 1995, the Massachusetts Water Resource Auth- before and after the start-up of the outfall pipe in the ority (MWRA) began an intensive monitoring programme Massachusetts Bay. designed to detect sources of fecal contamination in the Several previous studies have documented the presence of Massachusetts Bay, Boston Harbor, and its tributaries. The enteroviruses and adenoviruses in seawater (Tsai et al. 1993; primary objective of this research was to assess the presence of Girones et al. 1993; Abbaszadegan et al. 1993; Puig et al. 1994; anthropogenic viruses and their indicators (coliphages and Enriquez et al. 1995; Enriquez & Gerba 1995). Enteroviruses coliforms) in Massachusetts Bay. The secondary objective was (poliovirus, coxsackie virus types A and B, echoviruses) can to evaluate the impact of the Deer Island Sewage Treatment cause gastroenteritis, myocarditis and aseptic meningitis Plant’s outfall pipe on the occurrence of anthropogenic viruses (Melnick 1990). Adenoviruses type 40 and 41 can also cause and viral indicators. The outfall pipe was designed to increase gastroenteritis, but their presence in seawater is greatly the level of disinfection while reducing the level of chlorination underestimated because of difﬁculty isolating them in cell through a longer contact time. Because there is a poor culture. Enteric viruses survive longer than indicator bacteria understanding of how viruses and their indicators are inacti- in seawater (Melnick & Gerba 1980) and adenoviruses have vated during secondary treatment and disinfection, enteric been shown to survive longer than most enteric viruses in viruses, coliphages and indicator bacteria were monitored seawater (Enriquez et al. 1995). Indicator bacteria may be 60 Nicola A. Ballester et al. | Coliphages and anthropogenic viruses in Massachusetts Bay Journal of Water and Health | 03.1 | 2005

undetectable in a few days whereas enteric viruses may persist (TCVA-MPN) (USEPA 1995). This method required for several months (Wheeler 1990). Indicator bacteria (fecal samples to be inoculated onto buffalo green monkey kidney coliforms and Enterococcus) were studied to achieve a better cells (BGMK) and then evaluated for virus by visualization understanding of their role and suspected inadequacy as sole of cytopathic effects (CPE). However, other work done indicators of fecal pollution or as indicators of human viruses. from our laboratory (Chapron et al. 2000) previously The detection and presence of coliphages was also assessed for demonstrated that the level of viral contamination was their potential use as an additional indicator for fecal greatly underestimated when using the BGMK cell line contamination and viral presence. alone. Several enteric viruses do not exhibit CPE during Five sample sites in Massachusetts Bay were chosen based their replication cycle, while others such as astrovirus and on their proximity to the outfall pipe (Figure 1). Two sample rotavirus cannot replicate in this cell line. Both adenovirus sites were located on either side of the outfall diffuser, one and astrovirus require the addition of a proteolytic enzyme directly east and the other directly west of the diffuser head. for infection to occur. Hence, many of the epidemiologi- Shore sites were chosen near the coastlines northwest and cally important enteric viruses go largely undetected when southwest of the diffuser. One of shore sites was of particular using only the TCVA-MPN method. interest due to its use as a shellfish resource. The last site was Integrated cell culture nested polymerase chain reaction chosen in the mouth of the bay to the far northeast of the diffuser (ICC-nPCR) assay incorporates a cell culture step prior to head and acted as a control not influenced by sewage discharge. viral detection by PCR followed by nested PCR. The These five sites were sampled bimonthly over 7 years to study incorporation of a cell culture step permits viral replication the variability of anthropogenic contamination in the bay. resulting in an increase in the number of target nucleic acid From 1995 to 1999 enteric viruses were detected and copies (Pinto et al. 1995; Chapron et al. 2000; Reynolds et al. enumerated by the total culturable virus MPN assay 2001). The cell culture step also reduces the amount of

Figure 1 | MWRA outfall monitoring/sampling stations. 61 Nicola A. Ballester et al. | Coliphages and anthropogenic viruses in Massachusetts Bay Journal of Water and Health | 03.1 | 2005

inhibition typically seen in molecular techniques used with MATERIALS AND METHODS environmental samples. ICC-nPCR amplifies target viral nucleic acid sequences with the reverse transcriptase poly- TCVA-MPN method merase chain reaction (RT-PCR) for RNA viruses or PCR for During 1998–1999 water samples of 114–151 litres ( ø 30–40 DNA viruses followed by a nested polymerase chain reaction gallons) were collected using Zeta Plus MW (Cuno, Inc. (nPCR). The incorporation of nPCR into the assay increases Meriden, Connecticut) micro wound filters (Standard Methods sensitivity and specificity due to the use of primers internal to 1998). These samples were then tested using the TCVA-MPN the RT-PCR or PCR products, thus enabling detection of very method (USEPA 1995; Standard Methods 1998). Viruses were low numbers of specific viral particles. ICC-nPCR used with eluted from the filters with a 1.5% beef extract solution (pH 9.5, two cell lines (BGMK and CaCo-2) and the addition of a BBL Sparks, MD beef extract powder, 0.375% glycine). Eluates proteolytic enzyme may be a useful technique for the were concentrated by organic flocculation (pH 3.5) followed by detection and confirmation of a wide variety of enteric centrifugation (USEPA 1995). The pellet was resuspended with viruses in environmental samples. Additionally the cell sodium phosphate buffer (0.15 M Na2HPO4, pH 9.5), centri- culture step provides a means for infectivity testing. With fuged and the supernatant was adjusted to pH 7 for archiving direct PCR, infectious and non-infectious viruses could be and analysis. Each sample concentrate was passed through a detected whereas with ICC-nPCR the infectious nature of the beef extract treated 0.22-mm syringe filter to remove any viruses can be determined by comparing viral levels in the cell microbial contaminants prior to inoculation on BGMK cells. lysates compared with the concentrate. Four, 3 ml portions of filtered sample concentrate were each Routine monitoring could be an attractive alternative to inoculated onto 75 cm2 flasks of confluent BGMK cells. Flasks the costly and labour-intensive ICC-nPCR method if a good were incubated for 90 minutes at 378C with rocking every 15 indicator organism could be identified. The detection and minutes. Fifteen ml of serum-free maintenance cell culture evaluation of coliphages was important in this study media was added to each flask after incubation. Flasks were because of their potential use as indicators of fecal incubated at 378C and examined daily for CPE and cytotoxicity contamination and other anthropogenic viruses. Because for the first three days and then every other day for a total of 14 they have a similar structure and size to some enteric days. At the end of 14 days flasks were freeze thawed and 10% of viruses, coliphages have been suggested as possible indi- the first passage was put onto a new cell culture flask of cators of enteric viruses and may also aid in the detection of confluent BGMK cells for a second passage. Flasks that fecal pollution. Coliphages were originally detected using exhibited CPE were scored and the MPN/litre calculated. the single agar overlay method (USEPA 2001). This method, while being able to detect both male specific and somatic phages, only utilized a very small portion of the water Single agar overlay method sample. More recently, the modified two-step enrichment procedure (USEPA 2000) has been used to detect male During the same period, coliphage analysis was done on 1-l specific and somatic phages. The two-step enrichment grab samples by the single agar overlay method (USEPA procedure is desirable because it utilizes a larger sample 2001) and indicator bacteria (fecal coliforms and Entero- volume, vastly increasing its sensitivity to phage detection. coccus) were analysed by membrane filtration (Standard While not only increasing sample volume, it is thought that Methods 1998). Six 100 ml portions of water sample were the enrichment method is more sensitive to phage detection analysed by the single agar overlay method. Three portions because of an initial incubation step, in which low numbers were used for male specific coliphage detection and three of target coliphages are allowed to replicate. for somatic coliphage detection. One hundred ml of water In this study we looked at the correlations between the sample with 0.5 ml 4 M magnesium chloride hexahydrate detected levels of coliphages, indicator bacteria and human (MgCl2-6H2O) was warmed to 378C. Ten ml of log phase enteric viruses in relationship to seasonal variation and bacterial host (E.coli CN-13 or Famp) was added and mixed. proximity to the diffuser. The sample was warmed to 438C and added to 100 ml of 62 Nicola A. Ballester et al. | Coliphages and anthropogenic viruses in Massachusetts Bay Journal of Water and Health | 03.1 | 2005

tryptic soy agar (80 g l21). One ml of host appropriate reaction mixture was then added and run at 958Cfor5minutes, antibiotic (0.04 M nalidixic acid or a mixture of 0.001 M taq polymerase was added, and then subjected to 35 cycles of streptomycin and 0.004 M ampicillin) was then added, 958C, 30 s, 558C, 30 s, 728C, 30 s. Final extension was achieved mixed and poured evenly in Petri dishes. Plates were at 728C for 5 minutes. For nested PCR, 1 ml from each RT-PCR incubated for 24 hours at 378C and observed for plaques. reaction was added to a new tube containing 90 ml of a nested PCR reaction mixture which contained the primers 50- TCCGGCCCCTGAATGCGGCTA-30 and 50-GAAACACG- ICC-nPCR method GACACCCAAAGTA-30. Samples were run for 35 cycles of All samples after 1999 were evaluated by ICC-nPCR. Four-litre 958C, 30 s, 558C, 30 s, 728C, 30 s, yielding a 138 bp amplicon. grab samples were concentrated by mixing 40 grams of beef Twelve ml of each nested PCR product was run and sized on extract powder into the sample. Samples were brought to pH 1.8% agarose gels and stained with ethidium bromide. 3.5 and mixed for 30 minutes and centrifuged. Pellets were Molecular weights were determined by comparison with a 1 Kb DNA ladder (Life Technologies). Poliovirus LsC-1-2ab resuspended in 20 ml of 0.15 M Na2HPO4 (pH 9.5) buffer and centrifuged again. The supernatant was then adjusted to pH 7. was used as a positive control. Concentrates were filtered through beef extract pre-treated 0.22 mm filters prior to inoculation onto BGMK and CaCo-2 cells (Chapron et al. 2000). Two 3 ml portions of each sample Adenovirus PCR/nPCR concentrate were incubated for 30 minutes at 378Cwith The primers used were specific for adenovirus type 40 and 41. 21 21 5 mgml or 10 mgml of trypsin (Sigma St Louis, Missouri). Changes to the procedure described above included omission of 21 Samples containing 5 mgml were inoculated onto CaCo-2 the RT step and the primers (50-GCCGCAGTGGTCTTACATG- 21 cells and 10 mgml onto BGMK cells. The flasks were CACATC-30)and(50-CAGCACGCCGCGGATGTCAAAGT- incubated for 90 minutes at 378Cwithrockingevery15 30)(Puig et al. 1994). A 10-ml sample of cell lysate was denatured 21 minutes. Trypsin concentrations of 5 mgml were used for at 998C for 8 min. A 90-ml (final volume) PCR mixture was added 21 astrovirus and rotavirus and 10 mgml for adenovirus and to the denatured sample. The PCR parameters were the same as enteroviruses. Following incubation, 15 ml of serum-free described above. The nested procedure used was the procedure media was added to each flask. The flasks were incubated for described above. The primers utilized were (50-GCCACC- 5 days at 378C. After 5 days flasks were freeze thawed and cell GAGACGTACTTCAGCCTG-30)and(50-TTGTACGAGTAC- lysates were pooled. The cell lysates were analysed using the GCGGTATCCTCGCGGTC-3). These primers yield a 142-bp ICC/nPCR procedure for enteroviruses, astrovirus, rotavirus amplicon. Adenovirus type 40 and 41 were used as positive and adenovirus 40 and 41 (Chapron et al. 2000; Reynolds et al. controls. 2001). The laboratory procedures that were followed have been previously published (Chapron et al. 2000); changes to the methods are described below: Astrovirus RT-PCR/nPCR

The primers used were speciﬁc for human astrovirus, RT primer 50-GTAAGATTCCCAGATTGGT-30 and PCR primer Enterovirus RT-PCR/nPCR 50-CCTGCCCCGAGAACAACCAAG-30. A 10-ml sample of Enterovirus RNA was detected by RT-PCR using an RT primer cell lysate was denatured with 0.5 ml each of 0.05 M EDTA and (50-ACCGGATGGCCAATCCAA-30) and a PCR primer (50- downstream primer at 998C for 8 min. The RT mixture was CCTCCGGCCCCTGAATC-30)(Puig et al. 1994). A 10 ml added and run for 42 min at 428C and 5 min at 958C. After sample of cell lysate and denature reaction mixture was run at addition of PCR mixture the parameters were 958C, 5 min hot 998C for 8 minutes and then placed on ice. The reverse start, followed by 35 cycles of 958C, 30 s, 568C, 30 s, 728C, 30 s, transcriptase (RT) mixture was added and run for 42 min at with a ﬁnal extension at 728C for 5 minutes. For nPCR, the 428C and 5 min at 958C. The polymerase chain reaction (PCR) procedure was the same as described above, but the primers 63 Nicola A. Ballester et al. | Coliphages and anthropogenic viruses in Massachusetts Bay Journal of Water and Health | 03.1 | 2005

used were 50-CCTTGCCCCGAGCCAGAA-30 and 50- RESULTS TTGTTGCCATAAGTTTGTGAATA-30. These primers yield a 143 and/or 183-bp amplicon. Astrovirus serotype 2 was used Overall occurrence at all sampling sites of enteric viruses as a positive control. detected by ICC-nPCR remained relatively consistent from 1998 to 2002. In the samples from 1998–1999 that were reanalysed by ICC-nPCR, 36 (46%) of 78 were positive for one Rotavirus RT-PCR/nPCR or more enteric viruses and in the samples from 2000–2002, 55 (57.3%) of 96 samples were positive for one or more enteric The primers used were specific for rotavirus WA strain. The viruses. Viral detection of the same 1998–1999 samples when procedurewasthesameastheastrovirusRTPCR/nPCR done with the TCVA-MPN method demonstrated only 18 with the primers, RT primer 50-ATAGAAGACAGCGC- (23%) of 78 samples as positive. During this 5-year period the ACCGGATTTG-30 and PCR primer 50-ACAGACTTT- presence of adenovirus and astrovirus increased while CATTTGCGTCCGCAA-30. The PCR parameters were 958C, rotavirus and enterovirus presence decreased (Figure 2). 5 min hot start followed by 35 cycles of 958C, 30 s, 528C, 30 s, Several of the 1998–1999 samples that were originally positive 728C, 30 s, with a final extension at 728Cfor5minutes.The by the TCVA-MPN method (14.75%) were negative when nPCR procedure used the primers, 50-GACGCATCA- reanalysed by ICC-nPCR and several that were originally ACTGAAATAATAAAC-30 and 50-TGCACCAGCGAACATA- negative by the TCVA-MPN method (36.1%) were now CAGC-30. These primers yielded a 300-bp amplicon. Rotavirus positive by ICC-nPCR. WA strain was used as a positive control. Coliphage detection changed significantly from 1998 to 2000 due to the use of the more sensitive two-step enrichment Two-step enrichment method method. Detection of the male specific coliphages changed from 8.0% with the single agar overlay method to 58% with the Coliphage analysis was performed on 1-l grab samples by two-step enrichment and somatic coliphages from 9.8% to the modified two-step enrichment method (USEPA 2000) 55%. Indicator bacteria remained below statistically signifi- and indicator bacteria by membrane filtration (Standard cant counts (,30 cfu per plate) throughout 1998–1999 and Methods 1998). Aliquots of 500 ml of the water samples 2000–2002 in the Massachusetts Bay. were analysed for male specific and somatic coliphages. Detection method comparisons showed that the two- Twenty-five ml of tryptic soy broth (300 g l21), 6.25 ml 4 M step enrichment method was more sensitive than the single MgCl2-6H2O, 3 ml log phase bacterial host, and 5 ml host agar overlay method and that the ICC-nPCR method was specific antibiotic (0.04 M nalidixic acid or a mixture of more sensitive and specific than the TCVA-MPN method 0.001 M streptomycin and 0.004 M ampicillin) were added (Figure 3). Therefore the single agar overlay method and the to the samples and mixed. Samples were incubated for TCVA-MPN methods were not used after 1999. 24 hours at 378C. Between 10 and 20 ml of sample was then Correlations between the presence of coliphages, enteric spotted onto pre-poured tryptic soy agar plates containing viruses and indicator bacteria were based on presence, log phase host bacterium. Spot plates were incubated for proximity to the outfall pipe diffuser and seasonal variations. 24 hours at 378C and examined for lysis zones on the The presence of enteric viruses was significantly correlated bacterial layer. Roughly 50% of the lysis zones were plucked with the presence of male specific coliphages (r ¼ 0.682) and and reconfirmed by enrichment followed by spot plate. somatic coliphages (r ¼ 0.573), but not significantly correlated with the presence of indicator bacteria (r ¼ 0). Both male specific and somatic coliphages presence were significantly Statistical analysis correlated to one another (r ¼ 0.891) and to the presence of Pearson linear correlation was used to analyse the relation- adenovirus (r ¼ 0.651 and r ¼ 0.672 respectively), but not ships between organism presence, proximity to the outfall significantly correlated to the presence of astrovirus (r ¼ 0.122 and seasonal variation. and r ¼ 0.060 respectively). Male specific coliphages were also 64 Nicola A. Ballester et al. | Coliphages and anthropogenic viruses in Massachusetts Bay Journal of Water and Health | 03.1 | 2005

Data from 1998–1999 for coliphages obtained by the single agar overlay method and total enteric viruses by the TCVA-MPN method were done with the best available technology but should not be directly compared with the newer methods due to the lack of sensitivity of these assays. Reanalysis of these samples with ICC-nPCR demonstrated that the prevalence of total enteric viruses was highest in the spring. Specifically adenovirus, enterovirus and astrovirus were most prevalent in the spring, and rotavirus in the winter and summer. Adenovirus was not detected in the winter. Astrovirus and rotavirus were not detected in the autumn (Figure 4). Prevalence due to proximity to the outfall pipe diffuser Figure 2 | Changes in viruses present in the Bay. remained relatively similar from 1998–1999 to 2000–2002. In 1998–1999 male specific and somatic coliphages ¼ significantly correlated to the presence of rotavirus (r 0.692) analysed by the single agar overlay method were most ¼ and enterovirus (r 0.608), whereas somatic coliphages did prevalent at the shores northwest and southwest of the ¼ not have a strong significant correlation (r 0.504 and diffuser. Adenovirus, enterovirus and astrovirus were most ¼ r 0.419, respectively). prevalent directly to the west of the diffuser and rotavirus Between 2000 and 2002, seasonal levels of male specific directly to the east of the diffuser (Figure 5). In 2000–2002 coliphages and total enteric viruses were greatest in the male specific and somatic coliphages analysed by the autumn, whereas somatic coliphages concentration was enrichment method were most prevalent at both shore highest in the spring and winter. Specifically rotavirus was sites and directly to the west of the diffuser. Adenovirus was most prevalent in the spring, enterovirus in the summer, mostly prevalent directly east of the diffuser, rotavirus adenovirus in the summer and autumn, and astrovirus in directly to the west of the diffuser, astrovirus at the shore the autumn. Of interest is the absence of adenovirus and southwest of the diffuser, and enterovirus at the farthest site enterovirus in the spring and rotavirus in both the summer in the mouth of the bay (Figure 5). and autumn (Figure 4).

DISCUSSION

In this study we demonstrated the usefulness and beneﬁts of the ICC-nPCR method compared with the TCVA-MPN method in marine samples. The results from reanalysis of samples with ICC-nPCR show a great underestimation of viral levels when using the TCVA-MPN method alone. Use of the ICC-nPCR assay with two cell lines indicated that almost twice as many samples were positive (46%) for one or more viruses compared with 23% by the TCVA-MPN method. Reasons for the difference in results between methods could be due to mixed viral populations. When environmental samples are inoculated onto cell lines it is likely that there is a competitive Figure 3 | Comparison of detection methods. advantage for some viral groups to reach receptor sites for viral 65 Nicola A. Ballester et al. | Coliphages and anthropogenic viruses in Massachusetts Bay Journal of Water and Health | 03.1 | 2005

propagation before others and so detection of specific virus may also play a large part in the variations between data levels may be altered or absent. from the TCVA-MPN method and ICC-nPCR method. The incorporation of nested PCR into this method The increased sensitivity of the two-step enrichment achieved the increased sensitivity and specificity necessary method for coliphage detection compared with the single to detect very low numbers of viruses. The ICC-nPCR agar overlay method was also demonstrated. In the two-step method also provided a greater degree of certainty that the enrichment procedure the target phage is propagated during samples were either negative or positive for specific viruses the overnight incubation step allowing for very low due to the specificity of the internal primers rather than the numbers of phage to be detected, thus making it more subjectivity involved with the recognition of viral CPE. Of sensitive than the single agar overlay method. The two-step the samples reanalysed 14.75% that were originally positive enrichment method also uses a more representative sample by TCVA-MPN method were negative when analysed by volume and is simpler and more cost effective than the ICC-nPCR. The TCVA-MPN method is not specific to any single agar overlay method. Only 8.2% of the samples one virus and detects all culturable viruses that exhibit CPE. analysed were positive for male-specific coliphages with the Therefore the discrepancies in the data may be due to the single agar overlay method whereas 58% of the same sites presence of reoviruses or other non-human mammalian were positive using the two-step enrichment method. viruses in the original sample. In contrast, 36.1% of the Similar results were found with somatic phage, 9.8% were samples that were negative by TCVA-MPN were found to positive with the single agar overlay method and 55% with be positive when reanalysed by ICC-nPCR. This is primarily the two-step enrichment method. The differences in detec- due to the addition of a second cell line (CaCo-2) and the tion are due to the use of a larger volume of sample and the addition of a proteolytic enzyme prior to the cell culture pre-incubation step in the enrichment procedure. The pre- step for optimal propagation of rotavirus and astrovirus. incubation step also allows for propagation of phage in a Three of the four viruses detected by ICC-nPCR need the liquid medium that may otherwise be restricted in growth by addition of a proteolytic enzyme to infect the cells; therefore the single agar overlay method’s semi-solid medium. these may have gone undetected with the TCVA-MPN Correlations between the presence of coliphages, method as well. Inhibition and non-CPE forming viruses enteric viruses and indicator bacteria were based on

Figure 4 | Seasonal variation between 1998–1999 and 2000–2002. 66 Nicola A. Ballester et al. | Coliphages and anthropogenic viruses in Massachusetts Bay Journal of Water and Health | 03.1 | 2005

Figure 5 | Proximity variations to outfall diffuser 1998–1999 and 2000–2002. proximity to the treatment plant and seasonal variations. correlated to the presence of indicator bacteria (Hughes The presence of total enteric viruses was significantly et al. 1992; Wyer et al. 1995; Muscillo et al. 1997; Vantarakis correlated with the presence of male specific coliphages & Papapetropoulou 1998) and additionally that the num- and somatic coliphages. Both male specific coliphages and bers of enteroviruses and indicator bacteria vary spatially somatic coliphages were significantly correlated to each and temporally (Wyer et al. 1995; Vantarakis et al. 1998). other and to the presence of adenovirus, but not signifi- The four viral groups (enterovirus, adenovirus, astrovirus cantly correlated to the presence of astrovirus. Male specific and rotavirus) detected showed a seasonal variation as well. coliphages were also significantly correlated to the presence Rotavirus was most prevalent in the spring, enterovirus in the of rotavirus and enterovirus. Indicator bacteria were not summer, adenovirus in the summer and autumn, and astro- significantly correlated to any of the virus or phage, further virus in the autumn. Of interest is the absence of adenovirus demonstrating that indicator bacteria should not be used as and enterovirus in the spring and rotavirus in both the summer an indicator of viral presence. Seasonal relationships in and autumn. These seasonal changes in presence may be due 2000–2002 demonstrated that detection frequency of male to temperature differences, rain events and incidence of specific coliphages and total enteric viruses were highest in infection in the community. Also the absence of viruses the autumn. Somatic coliphages were highest in the spring could be due to the competitive attachment and replication and winter. Coliphage presence through the seasons encountered with mixed viral populations. indicates that monitoring of both male specific and somatic Samples from 1998–1999 reanalysed with ICC-nPCR coliphages is needed for them to be considered good compared with the samples from 2000–2002 showed that indicators of fecal pollution. From seasonal and proximity prevalence of total enteric viruses was highest in the spring correlation data, it appears that coliphages are better compared with the autumn, respectively. The seasonal indicators of viral presence, male specific coliphages having differences could have been due to climate changes for the highest correlations. In fresh water studies male specific Massachusetts Bay area and sample age. Samples from coliphages have been found to be more useful indicators of 1998–1999 were held for 2–3 years at 2808C before enteric virus presence (Sobsey et al. 1995). reanalysis. Specific viral presence when the older samples Previous studies of coastal marine waters found that the were reanalysed showed changes as well. Adenovirus, presence of enteric viruses could not be meaningfully enterovirus and astrovirus were most prevalent in the 67 Nicola A. Ballester et al. | Coliphages and anthropogenic viruses in Massachusetts Bay Journal of Water and Health | 03.1 | 2005

spring, and rotavirus in the winter and summer. Adenovirus coliphage and adenovirus as well as the highest number of was not detected in the winter, and astrovirus and rotavirus positives for enteroviruses were detected at this site in the were not detected in the autumn. No seasonal relationships 2000–2002 samples. Their occurrence is thought to be can be concluded from the data for two reasons. First, the either a result of boats releasing fecal material into the water data demonstrates large variations in viral occurrence. or from water currents and tidal changes. These could be due to temperature differences, rain events, While indicator bacteria have been shown not to be snowmelt, climate changes and human impact. Human good indicators of viral presence, more research needs to impact probably plays a large part in the seasonal variation be done on male specific and somatic coliphages and their results. The bay is used extensively for recreation in the role as viral indicators. Further prevalence and occurrence spring, summer and autumn. Second, the effects of mixed data needs to be compiled to elucidate the value of viral populations encountered with environmental samples coliphages as viral indicators. Other work done in our are still poorly understood. With mixed viral cultures laboratory (unpublished data) would indicate that site- competitive growth rates and the ability for attachment specific factors that are not yet understood influence the depend largely on the number of viral particles and the reliability of coliphages as an indicator organism. Hence, growth conditions. until these factors have been further studied, use of phage Prevalence of coliphage due to proximity to the outfall as an indicator needs empirical demonstration, through diffuser remained relatively similar from 1998–1999 to monitoring, for a particular site. Additionally, reovirus, 2000–2002. Coliphages were most prevalent at the shores hepatitis A virus and other enteric viruses also need to be northwest and southwest of the diffuser and directly to the included to better establish their relationship with coli- west of the diffuser. Presence at the shores was expected phages. In previous studies, reoviruses have frequently because of recreational use. The site to the west of the been detected in seawater and suggested as a possible diffuser had coliphage positives as well as adenovirus, indicator of fecal pollution (Muscillo et al. 1994; Tani et al. astrovirus and rotavirus positives but the indicator bacteria 1995; Patti et al. 1996). But since reoviruses are more results were negative. This further demonstrates the limited common in the animal population, the relationship use of bacterial indicators as indications of fecal pollution. between reoviruses and coliphages, or their use as No prevalence to proximity correlations could be made due predictors of other anthropogenic viruses, is not currently to the erratic changes of viral presence. The 5-year span of well defined. this study may also not be long enough to see any cyclic events within the viral populations due to proximity and seasonal variation. ACKNOWLEDGEMENTS The prevalence of coliphages and adenovirus at the southwest shore site may be due to recreational use, and are This work was funded in part by the Massachusetts Water of particular interest because it is a shell fishing resource area. Resources Authority, which also provided field support and Indicator bacteria numbers were below statistical significance laboratory analysis of bacteria indicators. This paper rep- (,30 cfu per plate) for this site, again indicating their resents the opinions and conclusions of the authors and not inadequacy as indicators of fecal pollution. Coliform bacteria necessarily those of the MWRA. We would also like to thank are routinely used as indicators of fecal pollution in water Dr Mark Sobsey for allowing the use of E. coli Famp. quality analyses. Using solely indicator bacteria to assess water quality is not adequate because of their lack of survivability in REFERENCES the environment. Coliphages would be a much more useful indicator, especially when detected by the enrichment method. Abbaszadegan, M., Huber, M. S., Gerba, C. P. & Pepper, I. 1993 Detection of enteroviruses in groundwater with the polymerase The site in the mouth of the bay, not influenced by chain reaction. Appl. Environ. Microbiol. 59, 1318–1324. sewage discharge, was to be used as a control to compare Chapron, C. D., Ballester, N. A., Fontaine, J. H., Frades, C. N. & with the other sites. Male specific coliphage, somatic Margolin, A. B. 2000 The Detection of astrovirus, enterovirus 68 Nicola A. Ballester et al. | Coliphages and anthropogenic viruses in Massachusetts Bay Journal of Water and Health | 03.1 | 2005

and adenovirus Type 40 and 41 in surface waters collected Reynolds, K. A., Gerba, C. P., Abbaszadegan, M. & Pepper, I. L. and evaluated by the information collection rule and 2001 ICC/PCR detection of enteroviruses and hepatitis A virus integrated cell culture/nested PCR procedure. Appl. Environ. in environmental samples. Can. J. Microbiol./Rev. Can. Microbiol. 60, 2520–2525. Microbiol. 47, 153–157. Enriquez, C. E. & Gerba, C. P. 1995 Concentration of enteric Sobsey, M. D., Battigelli, D. A., Handzel, T. R. & Schwab, K. J. adenoviruses 40 from tap, sea, and wastewater. Wat. Res. 29, 1995 Male-specific coliphages as indicators of viral 2554–2560. contamination of drinking water. AWWA Research Enriquez, C. E., Hurst, C. J. & Gerba, C. P. 1995 Survival of the Foundation Publication #90684. AWWA Research Foundation enteric adenoviruses 40 and 41 in tap, sea, and wastewater. Publication, Denver, CO. Wat. Res. 29, 2548–2553. Standard Methods for the Examination of Water and Wastewater Girones, R., Allard, A., Wadell, G. & Jofre, J. 1993 Application of 20th Edition 1998 American Public Health Association/ PCR to the detection of adenoviruses in polluted waters. American Water Works Association/Water Environment Wat.Sci. Technol. 27, 235–241. Federation, Washington, D.C. Hughes, M. S., Coyle, P. V. & Connolly, J. H. 1992 Enteroviruses in Tani, N., Dohi, Y., Kurumatani, N. & Yonemasu, K. 1995 Seasonal recreational waters of Northern Ireland. Epidemiol. Infect. distribution of adenoviruses, enteroviruses and reoviruses in 108, 529–536. urban river water. Microbiol. Immunol. 39, 577–580. Melnick, J. L. 1990 Enteroviruses: polioviruses, coxsackieviruses, Tsai, Y.-L., Sobsey, M. D., Sangermano, L. R. & Palmer, C. 1993 A echoviruses, and newer enteroviruses. In: Virology, 2nd simple method of concentrating enteroviruses and Hepatitis A edition. (Fields, B. N., Knipe, D. M., Chanock, R. M., Melnick, virus from sewage and ocean water for rapid detection by J. L., Roizman, B. & Shop, R. E. eds). Raven Press, New York, reverse transcriptase-polymerase chain reaction. Appl. pp. 549–605. Environ. Microbiol. 59, 3488–3491. Melnick, J. L. & Gerba, C. P. 1980 The ecology of enteroviruses in USEPA 1995 Virus Monitoring Protocol for the Information natural waters. CRC Crit. Rev. Environ. Contr. 10, 65–93. Collection Requirements Rule. US Environmental Protection Muscillo, M., Aulicino, F. A., Patti, A. M., Orsini, P., Volterra, L. & Agency, publication EPA/814-B-95-002. Government Printing Fara, G. M. 1994 Molecular Techniques for the identification Office, Cincinnati, Ohio. of enteric viruses in marine water. Wat. Res. 28,1–7. USEPA 2000 Method 1601: Male-specific (F þ ) and Somatic Muscillo, M., Carducci, A., La Rosa, G., Cantiani, L. & Marianelli, Coliphage in Water by Two-step Enrichment Procedure, draft C. 1997 Enteric virus detection in Adriatic seawater by cell April 2000. Office of Water, Washington, D.C. culture, polymerase chain reaction and polyacrylamide gel USEPA 2001 Method 1602: Male-specific (F þ ) and Somatic electrophoresis. Wat. Res. 31, 1980–1984. Coliphage in Water by Single Agar Layer (SAL) Procedure, Patti, A. M., Aulicino, F. A., Santi, A. L., Muscillo, M., Orsini, P., Draft January 2001. Office of Water, Washington, D.C. Bellucci, C., La Rosa, G., Mastroeni, I. & Volterra, L. 1996 Vantarakis, A. C. & Papapetropoulou, M. 1998 Detection of Enteric virus pollution of Tyrrhenian areas. Wat. Air Soil enteroviruses and adenoviruses in coastal waters of SW Pollut. 29, 1309–1316. Greece by nested polymerase chain reaction. Wat. Res. 32, Pinto, R. M., Gajardo, R., Abad, F. X. & Bosch, A. 1995 Detection 2365–2372. of fastidious infectious enteric viruses in water. Environ. Sci. Wheeler, D. 1990 The real risks of bathing in water contaminated Technol. 29, 2636–2638. by sewage. Environ. Health 98(10), 285–287. Puig, M., Jofre, J., Lucena, F., Allard, A., Wadell, G. & Girones, R. Wyer, M. D., Fleischer, J. M., Cough, J., Kay, D. & Merret, H. 1995 1994 Detection of adenoviruses and enteroviruses in polluted An investigation into parametric relationships between waters by nested PCR amplification. Appl. Environ. Microbiol. enteroviruses and faecal indicator organisms in the coastal 60, 2963–2970. water of England and Wales. Wat. Res. 29, 1863–1868.